Xavier b79ce29232 Add auditing instrumentation for python-oneviewclient

This change is about adding the ability of python-oneviewclient to
register time, parameters values and return values for methods
calls that do requests to OneView appliance. The instrumentation
implemented by this patch provides the necessary support that will
be needed to measure the performance of the OneView driver for
ironic and other tools.

Change-Id: Iae7fe8ec559ccb2cbf8f93486eb4a1a674d7053a
Co-Authored-By: Hugo Nicodemos <nicodemos@lsd.ufcg.edu.br>

2016-09-07 20:58:10 +00:00

2.8 KiB

Raw Permalink Blame History

python-oneviewclient

Library to use HPE OneView to provide nodes for Ironic

This library adds a communication layer between Ironic and OneView and abstracts the version of OneView in place.

Free software: Apache license
Documentation: http://docs.openstack.org/developer/python-oneviewclient
Source: http://git.openstack.org/cgit/openstack/python-oneviewclient
Bugs: http://bugs.launchpad.net/python-oneviewclient

Features

Audit logging

python-oneviewclient is capable of logging method calls to OneView for auditing. Currently, data about request timing and method names, parameters and return values, can be recorded to be used in the auditing process to discover and better understand hotspots, bottlenecks and to measure how the user code and OneView integration performs.

Enabling audit logging

To enable audit logging, the user code has to set three parameters in the constructor of the client object. namely: audit_enabled, audit_map_file and audit_output_file. audit_map_file and audit_output_file must be filled with the absolute path to the audit map file and the audit output file.

The audit map file

The audit map file is composed of two sections, audit and cases. In the audit section there should be a case option where one, and just one, of the audit logging cases needs to be specified. The cases section needs to be filled with a name for a case followed by the methods that the user wants to audit logging. The methods that are allowed for the audit logging are those decorated by @auditing.audit in python-oneviewclient.

See an example of an audit map file:

[audit]

# Case to be audit logged from those declared in cases section.

case = case_number_one

[cases]

# Possible auditable case name followed by the audit loggable
# methods' names.

case_number_one = first_method,second_method,third_method
case_number_two = first_method,third_method,fifth_method

The audit output file

The result of the audit logging process is a JSON formatted file that can be used by auditors, operators and engineers to obtain valuable information about performance impacts of using python-oneviewclient to access OneView, and better understand possible hotspots and bottlenecks in the integration of the user code and OneView.

See an example of an audit output file:

{
    "method": "get_node_power_state",
    "client_instance_id": 140396067361488,
    "initial_time": "2016-08-29T17:32:01.403420",
    "end_time": "2016-08-29T17:32:01.439126",
    "is_ironic_request": true,
    "is_oneview_request": false,
    "ret": "Off"
}

2.8 KiB Raw Permalink Blame History